Takeoffs and Trade: A Sorted Effects Approach

ABSTRACT We use an augmented production function to empirically examine whether accelerated trade growth affects the probability of achieving a sustained increase in economic growth, which we refer to as takeoffs. We make several empirical contributions. First, we propose a robust causal relationship between trade and takeoffs by generating a time‐varying geographic instrument. Our takeoff equation considers a fixed effects Logit model controlling for incidental parameters bias (Cruz‐Gonzalez et al. 2017). Then, using the sorted effects method, we estimate every partial effect of the impact of trade on the probability of a takeoff, in addition to the average partial effect. We identify sub‐groups of countries and their characteristics based on whether they are the most (90th percentile) or least affected (10th percentile) groups. We find that a one percentage point increase in the growth rate of the exogenous trade flows produces, on average, a 4.5% increase in the probability of a takeoff. However, the partial effects vary between 1.5% and 8.9%. Further, 86% of the countries that experience the highest increase in takeoff probability are non‐high‐income countries, ranging from low income to upper middle income, while 80% of the countries that experience the smallest increase in takeoff probability are high‐income economies.